COHERENT INTEGRATION FOR PASSIVE RADAR BASED ON SPARSE REPRESENTATION AND ATOMIC NORM MINIMIZATION

Passive radar (PR) systems utilize available illuminators of opportunity as signal sources to perform detection, tracking, and imaging tasks. Increasing the coherent integration time is the main way to improve the detection ability of passive radar. Compared with classic coherent integration methods based on cross-correlation, sparse representation (SR) based methods have the advantage of reducing sidelobes. In this paper, we propose an improved SR-based coherent integration method, in which sparse reconstructions are performed separately in Doppler and range dimensions. In Doppler sparse reconstruction, a range migration correction factor is introduced to solve the range migration problem. In the sparse reconstruction of the range dimension, the continuous sparse reconstruction based on the atomic norm is applied to overcome the offgrid problem brought by the traditional SR-based methods. Numerical simulation results show that the proposed method is superior to existing SR-based PR processing methods.